Corner attack on exo- and endo-tricyclo[3.2.1.02,4]octane by deuteron

Corner Bromination of 2-Methyl-endo-tricyclo[3.2.1.0]octane and -oct-6-ene. Andrew Burritt, James M. Coxon, and Peter J. Steel , William B. Smith. The...
0 downloads 0 Views 1MB Size
J. Org. Chem. 1989, 54, 1383-1391 The analytical scattering factors of Cromer and Waberloafor neutral atoms were used throughout the analysis; both real (D f') and imaginary (i D f'") components of anomalous dispersion were inc1uded.lob

Acknowledgment. We wish to thank the National (10)Internationat Tables for X-Ray Crystallography; Kynoch Press: Birmingham, England, 1974; (a) pp 99-101; (b) pp 149-150.

1383

Institutes of Health (Grant CA-11890) for financial support of this work. Funds for the purchase of the Nicolet R3n/V diffractometer system were made available from the National Science Foundation under Grant CHE-8514495. Registry No. la, 118895-55-9;lb, 118895-56-0;IC, 118895-57-1; Id. 118895-58-2:le. 118895-59-3:2. 118895-60-6: 3a. 118895-61-7: 3b; 118920-50-6;'3c,' 118895-62-8;3d,118895-63-9;$, 118895-64-0; 4%118895-65-1;4b, 118895-66-2;4C, 118895-67-3;4 4 118895-684 ZnC12, 7646-85-7; (CH3),SiCN,7677-24-9.

Corner Attack on ex0 - and endo -Tricycl0[3.2.1.0~~~]octaneby Deuteron and Mercuric Ions James M. Coxon,* Peter J. Steel,* and Barry I. Whittington Department of Chemistry, University of Canterbury, Christchurch, New Zealand

Merle A. Battiste University of Florida, Gainesville, Florida 3261 1 Received August 9, 1988

Reaction of endo-tricyclo[3.2.1.02~4]octane (1) with deuteron in methanol-dl gave a mixture (62:38) of methoxy ethers 3b and 4b from cleavage of the most substituted cyclopropane bond. The reaction proceeds by attack of deuterium at the corner of the cyclopropane. Trapping of the intermediate cation 5b is competitive with rearrangement to the nonclassical cation 6b. Reaction of exo-tricycle[ 3.2. 1.02v4]octane(2) under similar conditions gave methoxy ether l l b , which results from rupture of an external cyclopropyl bond, and methoxy ether 12b formed from rupture of the intemal cyclopropyl bond with inversion at the site of both electrophilic and nucleophilic attack. Reaction of hydrocarbon 1 with mercuric acetate in methanol gave 4-endo-(acetoxymercurio)-2-endometboxybicyclo[3.2.l]octane (34 from attack of the mercuric ion at the comer of the cyclopropane and nucleophilic attack by methanol with inversion without skeletal rearrangement. Similar reaction of the exo hydrocarbon 2 gave product 12c from intemal bond rupture without molecular rearrangement and with inversion of configuration at the site of electrophilic and nucleophilic attack. In addition an almost equal quantity of l l c , a product of rupture of the external cyclopropyl bond, was formed. The stereochemistry of mercuric ion and deuteron attack at C2 at the corner of the cyclopropane ring of 1 and 2 is rationalized by consideration of the symmetry and energy of the molecular orbitals involved. Acid- or electrophile-promoted ring cleavage of cyclopropanes has been the subject of considerable speculation.' The regiochemistry of reaction is often rationalized by a modified version of Markovnikov's rule, which states2that the ring opens between the carbons bearing the largest and smallest number of alkyl substituents. Thus, for cyclopropanes substituted at only one carbon the product(s) can be rationalized by ring opening leading t o the more substituted carbocation. However, with substituents on two carbons, products are generally observed to result from both Markovnikov-type addition and from cleavage of the most substituted cyclopropyl b0nd.l The two possible reaction trajectories for electrophilic attack on cyclopropane are shown in Scheme I. For substrates suitably labeled to allow the relative configurations a t C1 and C2 in the product to be established, these two processes can be differentiated. For example, DePuy3 found that cis1,2,3-trimethylcyclopropanereacted with deuteron to give 68% retention of configuration a t the site of deuteron (1) (a) Battiste, M. A,; Coxon, J. M. In The Chemistry of the CycEopropyl Group: Rappoport, 2.Ed.: J. Wiley and Sons: 1987. (b) Wiberg, K. B.; Kase, S. R. J.Am. Chem. SOC.1985,107,988 and 1005. (c) Lambert, J. B.; Schulz, W. J.; Mueller, P. H.; Kobayashi, K. J.Am. Chem. SOC. 1984,106,792. (d) Battiste, M. A.; Coxon, J. M. Tetrahedron Lett. 1986, 27, 517. (e) Kirmse, W.; Streu, J. J. Org. Chem. 1987,52, 515. (f)Coxon, J. M.; Steel, P. J.; Whittington, B. I.; Battiste, M. A. J. Am. Chem. SOC. 1988,110, 2988. (2) Zimmerman, M. P.; Li, H. T.; Dum, W. L.; Weeks, C. M.; Djerassi, C. J . Am. Chem. SOC.1984,106, 5602. (3) DePuy, C. H.; Funfschilling, P. C.; Andrist, A. H.; Olson, J. M. J.

Am. Chem. SOC.1977,99,6297.

Scheme I

(a) Corner attack by eleclrophiie (with inversion of configuration)

(b) Edge attack by eiectrophiie (with retention of configuration)

Scheme I1

attack and 32% inversion, as measured from the ratio of erythro and threo products. WibergIb has referred to retention (b, Scheme I) as cleavage of the C-C bond syn to the entering proton and to inversion (a, Scheme I) as cleavage of the C-C bond anti to the entering proton. Subsequent attack by the nucleophile in this system occurred with inversion.

0022-3263/89/1954-1383$01.50/0 0 1989 American Chemical Society

1384 J. Org. Chem., Vol. 54, No. 6, 1989

Coxon et al.

Scheme I11

Scheme IV

PtCI,

L

1

c1

Addition of bromine to cyclopropanes is slow in the absence of light unless the ring is heavily ~ubstituted.~In the presence of light, reaction is rapid with halogen radical attack on substituted cyclopropanes being fast even at -78 "C. The formation of HBr facilitates reaction by forming the conjugate acid of the cyclopropane. The more facile reaction of cyclopropane with HBr compared with bromine is in contrast to alkene chemistry where bromine addition is more rapid.Ia Nevertheless, cyclopropane does react with bromine and in an elegant study L a m b e r P showed that bromination of trans-cyclopropane-1,l ,2,3-d4(in an 8515 mixture with the cis isomer) reacted stereospecifically to give erythro product (85%),thereby establishing that inversion of stereochemistry occurred a t both the site of electrophilic and nucleophilic attack (Scheme 11). In this reaction electrophilic attack occurs a t the corner of the cyclopropane with inversion of configuration, and this contrasts with the less stereo~elective~ reaction of trimethylcyclopropane where attack at the edge with retention of configuration competes with reaction with inversion. Cyclopropanes can react with certain metals5 (e.g., PdCl,, PtC12,Rh,(CO),C12, and Ph3PIrC1CO)to undergo oxidative addition at the edge of the cyclopropane with the metal generally becoming attached to the most and least substituted carbon atoms of the ring. The first report of a reaction of a strained alkane with a metal complex was by TipperSa in 1955, who reported the reaction of cyclopropane and PtC1, (Scheme 111) to give a product, the structure of which was first suggested by Chatt and cow o r k e r ~in~1961 ~ to be a platinacyclobutane. An X-ray crystal structure by Mason and co-workersk confirmed the structure. For oxidative addition of metals to cyclop r o p a n e ~the , ~ ~metal must of necessity insert into the cyclopropane ring with retention of configuration at each of the carbons. Reactions involving the intermolecular insertion of an organometallic complex into an alkane C-C bond in a homogeneous reaction are rare since there is usually a substantial kinetic and thermodynamic barrier to C-C a ~ t i v a t i o n A . ~number ~ of examples are known of C-C bond cleavage in ligandse5g To date, no satisfactory explanation has been offered for the varied behavior of cyclopropanes with differing (4) Day, J. C.; Shea, K. J.; Skell, P . S. J . Am. Chem. SOC.1973, 95, 5089. (5) (a) Tipper, C. F. H. J. Chem. SOC.1955, 2045. (b) Adams, D. M.; Chatt, J.; Guy, R.; Sheppard, N. J . Chem. SOC.1961, 738. (c) Keeton, M.; Mason, R.; Russell, D. R. J. Organomet. Chem. 1971,33, 259. (d) Bailey, N. A.; Gillard, R. D.; Keetan, M.; Mason, R.; Russell, D. R. Chem. Commun. 1966, 396. (e) Crabtree, R. H. Chem. Reu: 1985, 85, 245. (f) Halpern, J. Acc. Chem. Res. 1982,15, 238. (9) Crabtree, R. H.; Dion, R. P.; Gibboni, D. J.; McGrath, D. V.; Holt, E. M. J . Am. Chem. SOC.1986,108, 7222. Kang, J. W.; Moseley, R.; Maitlis, D. M. J . Am. Chem. SOC.1969, 91, 5970. Benfield, F. W. C.; Green, M. L. H. J . Am. Chem. SOC.1969, 91, 5970. Benfield, F. W. C.; Green, M. L. H. J. Chem. SOC.,Dalton Trans. 1974, 1325. Eilbracht, P. Chem. Ber. 1976,109,1429,3136; 1980, 113,542, 1033, 1420, 2211. Grubbs, R. H., Myashita, A. J . Am. Chem. SOC.1978,100,1200,7418. Watson, P . L.; Roe, D. C. J. Am. Chem. SOC. 1982, 104,6471. Flood, T. C.; Statler, J. A. Organometallics 1984,3, 1795. Suggs, J. W.; Cox, S. D. J. Organomet. Chem. 1982,221,199. (h) Periana, 1986,108,7346. (i) Bishop, K. R. A.; Bergman, R. G. J. Am. Chem. SOC. C. Chem. Reu. 1976, 76,461. (j)Blomberg, M. R. A.; Siegbahn, P. E. M.; Backvall, J.-E. J. Am. Chem. SOC.1987, 109, 4450.

W H f E = D E = HgOAc

(3) ca 2 4 % 100 %

E E

= =

D 76 % HgOAc 0 %

(ai E

= H (b) E = D ( c ) E = HgOAc

electrophiles.'r2 There are two important aspects to this problem. The first is to understand when Markovnikov rupture of the cyclopropane is to be observed (i.e., regiochemistry) and the second the reasons for the varied trajectories found for attack of both electrophile and nucleophile (i.e., stereochemistry). For example, why is addition of a proton to 1,2,3-trimethylcyclopropanenot stereospecific3 while intermolecular nucleophilic attack on the resulting cation occurs exclusively with inversion? Why does the addition of bromine to cyclopropane occur with inversion at the sites of electrophilic and nucleophilic attack? Furthermore, oxidative addition of metals to cyclopropanes necessarily occurs with retention of configuration but may involve5hinitial insertion into a carbonhydrogen bond and subsequent rearrangement. Results and Discussion We now report the determination of the product(s) and stereochemistry of attack of proton, mercuric ion, and nucleophile in the reaction of endo- and exo-tricyclo[3.2.1.02~4]octane (1 and 2) in methanol. These compounds were selected for study because of the complimentary relationship of the cyclopropane ring to the skeleton allowing orbital effects to be differentiated from conformational and steric effects. endo-Tricyclo [3.2.1.0294]octane (1) was prepared by hydrogenation of endo-tricyclo[3.2.1.0z~4]~ct-6-ene6 while the exo isomer 2 was prepared by addition of methylene to n ~ r b o r n e n e . ~Methanol was chosen as solvent in preference to acetic acid, and therefore as nucleophile in the reaction with acid and mercuric ion, since an intermediate carbocation when trapped by methanol gives a product that is stable under the reaction conditions. In contrast, acetate products are often unstablelb and undergo further rearrangement, which complicates analysis of the primary reaction. Protonation of 1 in Methanol with H+ and D'. Reaction of 1 with methanol and a catalytic quantity of p-toluenesulfonic acidIf gives a high yield of 2-endomethoxybicyclo[3.2.1]octane (3a) when the reaction is carried out at 80 "C for 1 week. A GLC analysis of the reaction mixture indicated that the reaction had occurred to the extent of 80% with less than 2% of other products being formed. The axial nature of C2H in 3a, and hence the configuration of the methoxy group, was established (6) Closs, G . L.; Krantz, K. D. J. Org. Chem. 1966, 31, 638.

(7) Simmons, H. E.; Cairns, T. L.; Vladuchick, S. A.; Hoiness, C. M. Org. React. 1973, 20, 1.

J. Org. Chem., Vol. 54, No. 6, 1989 1385

Deuteron and Mercuric Ion Attack on Tricyclo[3.2.1.02~4]octane from the 'H NMR spectrum.* The 13C NMR spectrum could be assigned by comparison with the reported spectrum of 2-endo-hydroxybicyclo[3.2.l]octane~ and heteronuclear correlation experiments established connectivities.1° In order to establish the reaction trajectory of the proton in the reaction of endo-tricyclo[3.2.1.02~4]octane (1) with acidic methanol the reaction was carried out in methanol-d. The 13C NMR spectrum of the product showed deuterium incorporation a t both C4 and C6 (Scheme IV). From the intensities of the signals in the 2H NMR spectrum at 1.41 and 1.34 ppm, the ratio of 3b:4b was established as 62:38. The determination of the configuration of deuterium a t C4 and C6 for clarity is discussed separately below. The formation of the two deuterated products 3b and 4b can be accounted for if reaction of deuteron occurs exclusively at the corner of the cyclopropane with rupture of the most substituted adjacent cyclopropane bond. A protonated intermediate (5b) can be attacked by methanol to give 3b or collapse to the protonated species 6b, which will be attacked with inversion equally at both C1 and C2. A small isotope effect will perturb the symmetry of the cation 6b but is expected to be sufficiently small as to not effect the partition of the cation to 3b and 4b to any significant extent. The preferential attack on protonated hydrocarbon 1 by the nucleophile at C4 compared with C5 (62:38) demonstrates that protonated intermediate 5b formed by corner attack of deuterium with inversion is trapped by nucleophile to the extent of 24% before conversion to the symmetrical species 6b. In an independent control experiment, an authentic sample of 2-endo-methoxy-3,3,4-exo-trideuteriobicyclo[3.2.1]octane was shown to be stable under the reaction conditions, thus establishing the above ratio to be kinetic in origin. Capture of the unsymmetrical protonated cyclopropane 5b occurs before the activated complex has relaxed to any extent and is competitive with relaxation of the cation intermediate to the symmetrical nonclassical corner-protonated cation 6b.le Furthermore, the control experiment demonstrates that the reaction cannot proceed by exclusive capture of 5b followed by subsequent rearrangement. While nucleophilic attack on 5b occurs with inversion to give 3b, further intramolecular reorganization of the cation occurs to give 6b where the configuration a t C2 is retained. In this reaction it is apparent that intramolecular reaction with retention a t C4 and intermolecular nucleophilic attack a t C5 competes successfully with attack of methanol at C4. Assignment of the NMR Spectrum of 2-endoMethoxybicyclo[3.2.1]octane. The assignment of the chemical shifts of the C4 protons in P-endo-methoxybicyclo[3.2.l]octane was not possible from the chemical shift differences or from difference NOE studies and was determined in the following manner. 2-exo-Hydroxybicyclo[3.2.l]heptane was prepared by the method reported by LaLondell and a heteronuclear correlation experiment identified the connectivity of the C4 protons centered at 1.65 and 1.26 ppm with that carbon (28.3 ppm). (8) J = 10.1Hz C2H to C3-endo-H, J = 5.7 Hz C2H to C3-exo-H,and J = 3.0 Hz to C1H (C1H was further coupled to C8aH (6 Hz) and to the C7-exo-H (6Hz)). (9) Belikova, N. A.; Bobyleva, A. A.; Kalinichenko, A. N.; Lippmaa, E.; Ordubadi, M. D.; Pehk, T.; Plate, A. F. Org. Magn. Reson. 1976,8, 74. (10) H1, 2.36 ppm/Cl, 38.6; H2, 3.18 ppm/C2, 81.9; H3-exo, 1.82 and H3-end0, 1.23 ppm/C3, 25.7; H4-exo 1.33 and H4-endo 1.41/C4, 31.0; H5, 2.12/C5, 34.4; H6-exo 1.65 and H6-endo 1.34/C6, 28.8; H7-exo 1.73 and H7-endo 1.50/C7, 23.9; and H8a 1.59 and H8s 1.28/C8, 37.5. (11) LaLonde, R. T.; Ding, J.-Y.; Tobias, M. A. J. Am. Chem. SOC. 1967,89, 6651. Jefford, C. W.; Mahajan, S.; Waslyn, J.; Waegell, B. J . Am. Chem. SOC.1965,87, 2183.

c6

2el

I

2-

1.6

H6exo

1.2

1.4

H4endo H4exO H6endo

Figure 1. Heteronuclear correlation spectrum of a mixture (6238) of 4-endo- and 6-endo-deutero-2-endo-methoxybicyclo[3.2. lloc-

tane. Scheme V

Er

/

D

b Dz THF Pd C D 2 0 NaOD H

_(l)LiAIHa

+Br

(2)NaNHz

GD

Me1

RO I D

D 191

10 a R = H

b R=Me

A difference NOE experiment showed that irradiation of the signal centered a t 1.92 and identified as H8s due to coupling with H8a resulted in enhancement of the signal at 1.65 ppm, which was therefore assigned to the C4-exoH.12 For further confirmation of this assignment, a deuterated analogue of 2-exo-hydroxybicyclo[3.2.1]heptane (7)13was prepared (Scheme V) by reaction of 3-bromo-2exo-hydroxybicyclo[3.2.1]hept-3-ene (8) with deuterium gas, sodium deuteroxide (1M) in deuterium oxide-tetrahydrofuran in the presence of palladium on carbon (10%) as catalyst. Oxidation of the alcohol 7 with chromium trioxide-deuterioacetic acid gave the ketone 9, shown by analysis of the mass spectrum to contain on average 2.73 deuterium atoms. Lithium aluminum hydride reduction of this ketone where the deuterium configuration a t C4 is known gave the endo alcohol 10a in a mixture (4:l) with the exo alcohol 7, which was converted to a similar mixture (4:l) of the 2-endo-methoxy compound (lob) along with the exo-methoxy epimer by reaction of sodium amidemethyl iodide. The chemical shifts of the C4 protons of 3a are centered a t 1.33 and 1.41 ppm, and for the deuterated substrate lob, with a known exo configuration of deuterium a t C4, the presence of a triplet at 1.41 in the proton dimension of a heteronuclear correlation experiment allows the identification of this proton as endo. By exclusion the C4-exo-H of 3a is centered at 1.33 ppm. The multiplicity in the heteronuclear correlation experiment of the signal at 1.41 as a triplet results from C4-exo-deuterium coupling and connectivity of C4 and the endo proton. This unambiguously establishes the assignment (12) The reported assignment of the 13C NMR spectrumg of this alcohol is incorrect and C4, C6, and C7 are reassigned from coupling data obtained fromthe deuterated analogue. (13) A signal at 1.26 ppm in the 'H NMR spectrum of 7 and a triplet at 27.7 ppm due to C4 in the 13CNMR spectrum demonstrate not only that this substrate contains deuterium at C4 but that the configuration of the proton at that carbon is endo. The 13CNMR spectrum showed C3 as a multiplet consistent with the presence of two deuterium atoms at this position. The mechanism for formation of 7 may involve either reduction of an intermediate epoxide or loss of HBr and deuterogenation of the resulting alkene.

1386 J. Org. Chem., Vol. 54, No. 6, 1989

of the proton spectrum of the C4 protons in compound 3a. A related procedure for determining deuterium stereochemistry that uses a heteronuclear correlation experiment has recently been reported.14 The stereochemistry of the deuterium at C6 (in 4b) was established as endo from its chemical shift. In the heteronuclear correlation spectrum of the deuterated product mixture (3b,4b),the connectivities of the protons at C6 with that carbon were observed at 1.65 and 1.34 ppm and the protons at C4 with that carbon in 3b a t 1.33 and 1.41 ppm. The presence of 4b in the mixture was demonstrated from the heteronuclear correlation spectrum, which exhibited a triplet at 1.65 in the carbon dimension for C6exo-H, the additional splitting as a 1:l:l triplet resulting from coupling of C6 to the endo deuterium (Figure 1). The endo configuration of the C6-D eliminated for this compound a signal a t 1.34 ppm in the proton dimension and the presence of deuterium at C6 results in the signal at 28.5 for the C6 appearing as a triplet due to carbondeuterium coupling. The presence of 3b in the mixture was similarly identified by a triplet in the carbon dimension at 30.3 ppm, resulting from carbon-deuterium coupling and connectivity of C4 and the C4-exo-H. The absence of the endo proton in 3b eliminates a signal at 1.41 and the signal at 1.33 results from deuterium-carbon coupling and connectivity of C4 and the exo proton. Methoxymercuration of 1. Reaction of endo-tricycl0[3.2.1.0~*~]octane (1) with mercuric acetate in anhydrous methanol a t room temperature gave as a single product 4-endo-(acetoxymercurio)-2-endo-methoxybicyclo[3.2.1]octane (3c). The identity of this product was established in the following way. Reduction of 3c with sodium amalgam in sodium hydroxide gave 2-endo-methoxybicyclo[3.2.l]octane (3a), identical with an authentic sample. The stereochemistry and position of the mercury in 3c is consistent with the observed 'H NMR couplings (selective decoupling), and 13C-lg9Hgcouplings.lsJ6 Reduction of 4-endo-(acetoxymercurio)-2-endo-methoxybicyclo[3.2.l]octane (3c) with sodium amalgam in sodium deuteroxide gave 4-endo-deuterio-2-endo-methoxybicyclo[3.2.l]octane (3b). The 2H NMR spectrum showed a signal at 1.43 ppm and a two-dimensional lH-13C heteronuclear correlation experiment identified the C4-exo-H at 1.35 ppm, thereby establishing the configuration of the deuterium." Reduction with sodium borodeuteride was not stereospecific and gave rise to 4-exo- and 4-endodeuterio-2-endo-methoxybicyclo[ 3.2.lloctane in the ratio of 45:55, respectively. Formation of 3c results from attack of the mercuric ion at the corner of the cyclopropane ring with cleavage of the internal cyclopropyl bond and subsequent attack by methanol with inversion. The unsymmetrical mercurated cation 5c does not rearrange to the more symmetrical mercurated cation 6c. A high degree of bond order is maintained between C4 and C2 in the cation 5c with little charge localization at C4. This reaction is therefore similar to that of alkenes with mercuric acetatela where skeletal rearrangement is not normally observed even though charge development on the alkene carbons is observed. (14) Reese, P. B.; Trimble, L. A.; Vederas, J. C. Can. J . Chem. 1986, 64, 1427. (15) Kitching, W.; Atkins, A. R.; Wickham, G.; Alberts, V. J. Org. Chem. 1981,46, 563. (16) J1~c-l~Hc = C1, 25; C2, 310; C3, 97; C4, 1634; C5, 66; C6, 65; C7, n.0.; C8, 310 Hz. (17) The reduction conditions have been shown in related systems16 to occur with retention of configuration and therefore the stereochemistry and position of the mercury in 2c is unambiguous. (18) Wardell, J. L. Comprehensive Organometallic Chemistry; Wilkinson, G., Stone, F. G. A., Abel, E. W., Eds.; 1982; Vol. 2, p 863.

Coxon et al. Scheme VI

Y

(a] E = H lbl E = D ( c ) E = HQOAC

. /A (141

(12)

E = D = HQOAC

E

76 O h 43%

Protonation of 2 in Methanol with H+ and D+. LaLonde some years studied the reaction of exotricyclo[3.2.1.02~4]~ctane (2) with sulfuric acid in acetic acid and observed by GLC five acetate products, two of which were identified after isolation and hydrolysis as 2-exobicyclo [ 3.2.11octanol (35% ) and 2-hydroxybicyclo [2.2.21 octane (22%). Both of these products result from cleavage of the internal cyclopropyl bond. We decided to reinvestigate the ring-opening reaction of this hydrocarbon with acidic methanol and to contrast its reaction with mercuric acetate in methanol, conditions that would be expected to give a less complex reaction mixture since the initial products are, under these conditions, expected to be stable. The reaction with methanol in the presence of p-toluenesulfonic acid was followed by GLC and gave 2-exo-methoxybicyclo[3.2.l]octane(12a) and 2-exo-methoxy-7-syn-methylbicyclo[2.2.l]heptane( l l a ) in 76% and 15% yield, respectively. These products were separated by preparative GLC. The identity of 2-exo-methoxybicyclo[3.2.l]octane (12a) follows from the 'H NMR and I3C NMR spectra and by its identity to an authentic sample prepared by reaction of 2-exo-hydroxybicyclo[3.2.l]heptane prepared as above (Scheme VI) with sodium amide-methyl iodide.20 The identity of 2-exo-methoxy7-syn-methylbicyclo[2.2.l]heptane(1 la) follows from a heteronuclear correlation experiment and difference NOE studies.21 The stereospecific formation of 3-exo-methoxybicyclo[3.2.l]octane (12a) from exo-tricyclo[3.2.1.O2v4]octane(2) and 3-endo-methoxybicyclo[ 3.2.lloctane (3a) from endo-tricyclo[3.2.1.02~4]~ctane (1) excludes the intermediacy of a common classical secondary cation from cleavage of the C2-C4 bond of the hydrocarbons since the stereochemistry of nucleophilic attack on such a cation would be independent of the configuration of the starting cyclopropyl ring. LaLondell and co-workers analyzed the stereochemistry of the incorporated deuterium in the major product, 2exo-acetoxybicyclo[3.2.l]octane,from reaction of hydrocarbon 2 with sulfuric acid-d, in acetic acid-dl. The acetate was converted into hydrocarbon and a comparison of the (19) LaLonde, R. T.; Batelka, J. J. Tetrahedron Lett. 1964, 445. (20) Kirmse, W.; Olbricht, T. Chem. Ber. 1975, 108, 2606. (21) The coupling of the C1H (2.10 ppm) to the C6-exo-H (3.1 Hz), and C4H (1.95 ppm) to C3-exo-H (3.0 Hz) and C5-exo-H (3.0 Hz), establishes that either the methyl or the methoxy group is at C2 and in an exo configuration. A difference NOE experiment showed that on irradiation of H4 (1.95 ppm) the signal of the methyl was enhanced as well as the C3Hz (1.73 ppm), C5-exo-H (1.5 ppm), and C5-endo-H (1.05 ppm). Irradiation of H2 and the methoxymethyl, both centred at 3.3 ppm, caused enhancement of H1 (2.1 ppm), C3Hz (1.73 ppm)! and the methyl at 1.05, thereby establishing that the methoxy group is at C2 and the methyl in a syn configuration at C7.

Deuteron and Mercuric Ion Attack on Tricyclo[3.2.1.02~4]octane infrared spectrum of the deuterated hydrocarbon with the infrared spectra of authentic exo- and endo-Zdeuteriobicyclo[3.2.l]octane made. This study concluded that the deuterium was exo, consistent with deuterium attack at the corner of the cyclopropyl unit. Our reinvestigation of the stereochemical aspects of the reaction of exo hydrocarbon 2 with methanol-d and deuteron were decidedly more straightforward as a consequence of advanced instrumental techniques. The mass spectrum of the product 12b showed the deuterium content of the sample to be 2% do, 90% dl, and 8% d2. A heteronuclear correlation experiment on undeuterated sample 12a established the C4-exo-H at 1.62 ppm and the C4-endo-H a t 1.38 ppm. For the deuterated sample a 2H NMR spectrum exhibited a signal for the C4-exo-D at 1.62 ppm and a heteronuclear correlation experiment exhibited a triplet in the carbon dimension exhibiting connectivity with a proton at 1.38 in the proton dimension. The splitting results from coupling of C4 to the exo deuterium. Furthermore, irradiation of H8s at 1.84 ppm caused no enhancement of the signal of the 4-endo-H at 1.38 ppm, further confirming the configuration of deuterium a t C4 as exo. The partition of the reaction of 2 to give a greater yield of 12b compared with l l b is either a result of protonation at C2 being more favored than protonation at C3 or alternatively that the C2-protonated species is more rapidly trapped than the C3-protonated species. Failure to observe deuterium at C6 or formation of the endo-methoxy isomer of 12 from the reaction with D+ methanol is consistent with nucleophilic attack on the C2-protonated species occurring before the cation has relaxed to allow conformational change and overlap of the developing cation center with the Cl-CB u bond. The presence of deuterium in the methyl of l l b was established from the 13C NMR spectrum, which showed the methyl carbon as a 1:l:ltriplet. Methoxymercuration of 2. Reaction of the hydrocarbon 2 with mercuric acetate in water followed by reduction of the intermediate hydroxy acetoxymercurio adducts has been reportedz2to give a 1:9 mixture of 15a and 16a, both products resulting from external bond rupture involving mercury attack at C3. The stereochemistry of

(15)

a E = H b E = D C E = HgOAc

(16) a E = H

b E = D C E = HgOAc

the alcohol in 15a was, however, undetermined. Oxidation of a mixture of 15a and 16s gave a similar mixture of the corresponding ketones 3-exo-methyl- and 7-syn-methylbicyclo[2.2.l]heptan-2-ones.When the reaction of 2 with mercuric acetate was repeated, but in methanol as solvent, 1lc and 12c were formed in approximately equal amounts. The mercuric compounds could not be separated and therefore were reduced to give the corresponding methoxy hydrocarbons l l a and 12a in 48% and 43% yields, respectively. These two products were separated by preparative GLC and shown to be identical in all respects with the products obtained from reaction of the hydrocarbon with methanol in the presence of p-toluenesulfonic acid. The observation of mercuric ion cleavage of the most substituted cyclopropane bond in the formation of 12c is in direct conflict with previously stated rules for mercuric ion induced cyclopropane ring opening.18

J. Org. Chem., Vol. 54, No. 6, 1989 1387 Scheme VI1

CN

CN

The mixture of mercuric salts from reaction of the exo hydrocarbon 2 with mercuric acetate in methanol was reduced with sodium amalgam in sodium deuteroxide. For the product l l b the presence of deuterium in the methyl syn to the methoxy group was established from the 13C NMR spectrum, which showed the methyl carbon as a 1:l:l triplet centered at 12.7 ppm. Notably a 6,2-hydride shift does not compete with the capture of the intermediate cation by methanol for this reaction or for reaction with a proton. This is consistent with delocalization from the C3-mercurated species to C1 and blocking C2 from exo face attack as shown in Scheme VI. The 2H NMR spectrum of 2-exo-methoxy-4-exo-deuteriobicyclo[3.2.l]octane showed a signal at 1.62 ppm and the product was identical in all respects with that obtained from the reaction with acidic methanol-d, thereby establishing that the mercury attacked the cyclopropyl ring with inversion. Regioselectivity and Stereochemistry of Reactions of 1 and 2. The regioselective cleavage of the C2-C4 bond of hydrocarbon 1 with acid and with mercuric ions in the formation of 3 and 4 (Scheme IV) reflects either (i) that an energy difference of the cations is such that there is a preference for protonation and mercuration at C2 (or C4) as compared with C3 or (ii) that the bond rupture of the C2 (or C4) protonated (mercurated) hydrocarbon is faster than bond rupture of the C3 protonated (mercurated) hydrocarbon. The reaction parallels that of endo-tricycl0[3.2.1.O~~~]oct-6-ene with acetic acidz3and with tetra~ y a n o e t h e n e(Scheme ~~ VII). In both reactions it is the more substituted bond that is cleaved. This compares with similar reaction of the exo hydrocarbon 2 where external bond cleavage successfully competes with rupture of the internal cyclopropyl bond (Scheme VI). This latter situation finds a parallel in the reaction of 1,1,2,2-tetramethylcyclopropane with acetic acid,lb which gives products that result from cleavage of the most substituted bond along with products from rupture of the adjacent bond. Wiberglb has calculated that for the reaction of the tetramethylcyclopropane with a proton there is not a strong preference for which bond is cleaved. He argues that the comparability in calculated energy of cations 17 and 18, CH3

(17)

A

(18)

which are presumed to lie along the reaction coordinate, rationalizes the observation that the electrophile does not have a strong preference for which bond is cleaved. Nevertheless these reactions always proceed so that the nucleophile or its equivalent, whether an intermolecular species such as an acetate ion, or intramolecular nucleophile such as a migrating group (H, Me) or an adjacent bond becomes associated with the more substituted cat(23) Battiste, M. A.; Coxon, J. M.; Edelman, R. Tetrahedron Lett. 1972, 4577.

(22) Reutov, 0. A.; Rodina, N. B.; Sokolov, V. I. J.Organomet. Chem. 1969, 17, 477.

(24) Battiste, M. A.; Coxon, J. M.; Posey, R. G.; King, R. W.; Mathew, M.; Palenik, G. J. J. Am. Chem. SOC.1975, 97, 945.

Coxon e t al.

1388 J . Org. Chem., Vol. 54, No. 6, 1989

es

e5

Figure 2. Mixing of the C1, C8/C5, C8

cyclopropane orbitals.

ea

a

orbitals with the

ionic center. While the reactions proceed toward the most stable carbocation, at the product-forming activated complex the structure is considered not to have relaxed to a large extent since relief of steric strain does not correlate with reaction rate.lb The products are considered to be formed by capture of the protonated cyclopropane before it has become an open carbocation, thereby resulting in stereoselective capture by nucleophile or stereoselective hydrogen migration or loss. A notable feature of the reactions of the endo hydrocarbon 1 is that the attack by an external nucleophile is always observed to occur with inversion of configuration. Polarization of the cyclopropane ring by an attacking electrophile has been found to be influenced by the lowlying u orbitals present in the molecule.1b For the endo hydrocarbon 1 external bond cleavage does not compete with cleavage of the internal bond. Overlap of the C2-C4 bond with the Cl-C8 and C5-C8 bonds will raise the energy of the cyclopropane HOMO’s with a C2-C4 component and thereby reduce the energy difference between these orbitals and the LUMO hydrogen 1s orbital favoring cleavage of this more substituted bond. It is noteworthy that for the exo isomer overlap of the C2-C4 bond with the Cl-C8 and C5-C8 bonds is not possible and the adjacent external bond cleaves in competition with the internal bond. Mixing of the Walsh 3e’(S) and 3e’(A) orbitals of endo-tricyclo[3.2.1.02~4]octane (1) with the low-lying C1C8/C5-C8 orbitals gives rise to the e, and e, molecular orbitals shown in Figure 2. From photoelectron spectroscopy, the e, orbital can be assigned as the HOMO at -9.40 eV, the e, orbital at -10.20 Attack at the e, orbital of necessity results in C2-C4 bond cleavage, and attack could occur with either retention or inversion. For edge attack, while there is a favorable interaction with the cyclopropyl3e’(S) component of e,, an unfavorable secondary orbital interaction with the Cl-CS/C5-C8 orbitals disfavors such attack. This edge attack is also disfavored by the unfavorable interaction of the proton 1s orbital with the e, orbital. For attack at the corner no large antibonding interaction is present, and hence this pathway is favored. As interaction of orbitals of similar energy produces the greatest orbital energy stabilization gain influencing transition-state energy, overlap of the electrophile LUMO with the energetically lower e, molecular orbital, which could give external bond cleavage, is not as important as overlap with the e, HOMO, and therefore cleavage of an external cyclopropyl bond is not observed. For hydrocarbon 2, mixing of the Walsh 3e’(S) and 3e’(A) orbitals with the Cl-C7/C5-C6 orbitals of the correct symmetry gives rise to the HOMO e, and the e, orbitals shown in Figure 3. The separation between the For attack e, and e, orbitals for 2 is not as large as for lez5 at the e, orbital, corner attack is more favored since edge (25) (a) Bischof, P.; Heilbronner, E.; Prinzbach, H.; Martin, H. D. Helu. Chim. Acta 1971,54, 1072. (b) Jonkers, G.; Van der Meer, W. J.; DeLange, C. A.; Baerends, E. J.; Stapersma, J.; Klumpp, G. W. J. Am. Chem. SOC.1984, 106, 587.

Ea Figure 3. Mixing of the 3e’(S) and 3e’(A) orbitals with the C1, C7/C5, and C6 u orbitals to give the e, HOMO and the e, mo-

lecular orbital.

attack at C2-C4 ‘isdisfavored by an unfavorable secondary orbital interaction with the Cl-C7/C5-C6 orbitals. External cyclopropyl bond cleavage attack at the e, orbital could result in either retention or inversion; however, determination of the stereochemistry of electrophile attack was not possible due to the presence of the prochiral C8 methylene in product 11. For proton attack at hydrocarbon 2, the preference for corner attack at the e, HOMO reflects the favorable stabilization to the transition state from this HOMO/LUMO interaction. However since the energy gap between e, and e, is not large, and delocalization stabilizes cation 13a produced from attack at the e, orbital, this pathway is competitive. For attack by mercuric acetate on 2, any steric hindrance to corner attack at the e, orbital is offset by the gains in the favorable e, HOMO/mercury 6s LUMO interaction. Attack by mercuric acetate at the e, orbital is, however, more competitive than proton attack at 2.26 The stereochemistry observed for proton and mercuric ion attack at C2(C4) of both the exo and endo hydrocarbons 1 and 2 parallels that found for attack of bromine on trans-1,1,2,3-tetradeuteriocyclopropanewhere the formation of erythro product excludes the intermediacy of a symmetrical corner-brominated cyclopropane interand dictates that the reaction pathway is unsymmetric. LamberP has interpreted the reaction to require a corner-brominated cyclopropane, which by definition is symmetrical. For this to be the case rotation of the pentacoordinated carbon would have to be slow with respect to the rate of nucleophilic attack. Furthermore, the initial configuration of the bromine atom in the species would have to dictate which of the adjacent carbon sites is attacked by nucleophile. The favored attack by the electrophiles deuteron and mercuric ion and bromine at the corner of the cyclopropane ring with inversion reflects in general the favorable interaction of both the degenerate HOMO’s of the cyclopropane with the hydrogen 1s or LUMO 6s of the mercuric ion or the 4p orbital of bromine respectively (Figure 4). For edge attack the HOMO/ LUMO interaction with the symmetric Walsh orbital is favorable for proton, mercuric ion, or bromine, but this is not the case with the unsymmetric orbital. A favorable interaction of the LUMO Walsh orbital of cyclopropane with the d orbitals of electron-donor metals allows oxidative addition2‘ at the edge of the cyclopropane. This interaction compensates for the more favored r~ interaction at the corner of cyclopropane between the HOMO Walsh orbitals and the LUMO orbitals of the electrophile. The low promotion energy generally observed for electron donation from the nd orbital to the (n + 1)p (26) Exclusive external cyclopropane ring cleavage has been reported“ for reaction of hydrocarbon 2 with mercuric acetate in water. The cationic nature of 14c is low with the positive charge residing at mercury. Cation 13c, from which rearrangement occurs, is therefore expected to be more sensitive to stabilization by a more polar solvent. (27) (a) Campbell, W. H.; Jennings, P. W. Organometallics 1983, 2, 1460. (b) Waddington, M. D.; Campbell, J. A,; Jennings, P. W.; Campana, C. F.; Organometallics 1983, 10, 1269. (c) Wiberg, K. B.; McClusky, J. V.; Schulte, G. K. Tetrahedron Lett. 1986, 27, 3083.

Deuteron and Mercuric Ion Attack on Tricyclo[3.2.1.02~4]~ctane

(a) IC

;

I

I

J. Org. Chem., Vol. 54, No. 6,1989 1389

[3.2.1.02~4]octane (0.90 g, 90%) as a low melting glass, mp 71-72 "C: 'H NMR 8H (CDC13) 2.21 (9, Wh/z = 7 Hz, H1, H5), 1.88 (m, 'J&,ga= 8.3 Hz, 3J&,~ = 3Jg,5 = 3J&,&ndo = 3 J ~ , 7 e n d 0= 2.1 Hz, H ~ s ) , 1.44 (dd, 2Jga,8s = 8.3 Hz, dJ8a,3exo = 2.4 Hz, H8a), 1.36-1.30 (m, H2, H4, HGexo, H7exo), 1.01 (dd, 'J&ndo,@Xo = 7.3 Hz, 3J&ndo,7exo = 2.3 Hz, HGendo, H7endo), 0.88 (dt, 'J%ndo,&xo = 5.9 Hz, 3Jhndo2 = 3J3end0,4 = 2.4 Hz, H3endo), 0.72 (m), 2J~xo,&ndo= 5.9 Hz, 3J3e.10,2 = 3J3exo,4 = 7.4 Hz, 5J3exo,8a = 2.3 Hz, H3exo); 13C NMR, see ref 29.

Corner

exo-Tricycl0[3.2.1.O~~~]octane (2). exo-Tricyc10[3.2.1.0~~~]octane was prepared from norbornene by addition of methylene carbene generated by using a zinc/copper ~ o u p l e :'H ~ NMR b~ (CDC13) 2.21 (9, Whp = 5.1 Hz, H1, H5), 1.43 (Wh/z = 18 Hz, HGexo, H7exo), 1.24 (Whiz = 13 Hz, HGendo, H'iendo), 0.92 (dt, 2J8s,ga= 10.4 Hz, 4Jgs,6endo = 4Jg8,7endo = 2.1 Hz, H ~ s )0.66 , (dd, 3Jz/4,3endo = 7.3 Hz, 3Jz/4,3exo = 3.2 Hz, H2, H4), 0.57 (d, 'Js,,~, = 10.4 Hz, H8a), 0.28 (dt, 2J3exo,3endo = 6.0 Hz, 3J3exo,2 = 3J3exo,4 = (b) edge 3.1 Hz, H3exo), -0.11 (m, 2J3endo,3exo = 6.1 Hz, 3J3endo,2 = 3J3endo,4 Figure 4. u interaction of LUMO of electrophile with degenerate = 7.0 Hz, H3endo); partial lH NMR, see ref 30; I3C NMR, see HOMO'S of cyclopropane. ref 29. Reaction of 1 with p -Toluenesulfonic Acid/Methanol. orbital5' indicates that t h e s e orbitals a r e energetically endo-Tricyclo[3.2.1.02~4]octane (120 mg), anhydrous methanol (3 similar. It is therefore valid t o consider interaction bemL), and p-toluenesulfonic acid monohydrate (8 mg) were placed tween the cyclopropyl3e' HOMO and the metal ( n l ) ~ in an ampule (5 mL) and kept a t 80 OC for 7 days. The mixture LUMO orbitals. F o r m e r c u r y the d o n o r abilityz8 of t h e was diluted with water ( 5 mL) and the product extracted into d, orbitals is small and thus t h e d, HOMO-cyclopropane pentane. The pentane extracts were washed with aqueous sodium LUMO interaction is u n i m p o r t a n t and t h e reaction stercarbonate solution and dried over MgS04, and the solvent was eochemistry parallels t h e reaction with deuteron. removed under reduced pressure to give an oil (130 mg, 81%) shown by GLC analysis to be a t least 95% pure and shown to be 2-endo-methoxybicyclo[3.2.l]octane (3a): 'H NMR bH (CDCl,) Conclusion . 3.30 (s, OMe), 3.18 (3Jk0,3ed0 = 10.1 Hz, 3 J ~ x 0 , = ~ 05.7 Hz, 3Jk0,1 T h e general preference for corner a t t a c k l f at cyclo= 3.0 Hz, H2exo), 2.36 (3J1,zexo = 3.0 Hz, 3J1,7exo = 6.0 Hz, 3J1,S, p r o p a n e by p r o t o n s and mercuric ions reflects the favor= 6.0 Hz, H l ) , 2.12 (s, w h / z = 7 Hz, H5), 1.9-1.2 (m, 10 H's). The able interaction for both degenerate HOMO orbitals of t h e following assignments were obtained from a heteronuclear corcyclopropane with t h e LUMO of t h e electrophile. For t h e relation experiment: 1.82, H3exo; 1.73, H7exo; 1.65, H6exo; 1.59, molecules 1 and 2 where the 3e'(S) and 3e'(A) Walsh orH8a; 1.50, H7endo; 1.41, H4endo; 1.34, H6endo; 1.33, H4exo; 1.28, H8s; 1.23, H3endo. I3C NMR: bc (CDCI,) 81.9 (C2), 55.7 (OMe), bitals a r e no longer degenerate (see Figures 2 and 3), t h e 38.6 (Cl), 37.5 (C8), 34.4 (C5), 31.0 (C4), 28.8 (C6), 25.7 (C3), 23.9 reactions w i t h electrophiles a r e m o s t influenced by t h e (C7). HOMO, since this interaction will give t h e greatest energy Preparation of 2-end0 -Methoxybicyclo[3.2.l]octane and gain u p o n interaction with t h e LUMO of the electrophile. the Deuterated Analogue (lob). A solution of 3-bromoT h e next lower occupied molecular orbital, in t h e absence bicyclo[3.2.l]oct-3-en-2-exo-ol (8)" (4 g) in tetrahydrofuran (30 of a n y antibonding interactions, becomes important as t h e mL) with aqueous sodium hydroxide (30 mL, 1 M) was shaken energy difference between this orbital and t h e HOMO with prehydrogenated palladium on carbon (2.4 g, 10%) under decreases. hydrogen at 1.5 atm. After the addition of hydrogen was complete, the mixture was diluted with water and saturated with NaCl, and General Methods the product was extracted with ether. The solvent was removed by distillation and the crude product steam distilled to give NMR spectra were obtained on a Varian T-60 or Varian XL-300 bicyclo[3.2.1]octan-2-exo-ol'1(1.88 g, 75%) as a white solid: 'H (300 MHz 'H, 75 MHz 13C,46 MHz 2H) instrument. All 2H NMR NMR bH (CDC13) 3.73 (s, Whp = 8 Hz, H2), 2.21 (Hl), 2.16 (H5), spectra were run unlocked with broadband proton decoupling, 1.92 (d, 'JgS,8,= 10.0 Hz, H&), 1.75 (s, Wh/Z = 3 Hz, OH), 1.69 an acquisition time of 4 s and using 2 drops of CDC1, (7.27 ppm) (m, H4ex0, HGexo, H7ex0, HSexo), ca. 1.5 (H7endo), 1.39 (m, as an internal reference. Heteronuclear proton-carbon correlation HGendo, H3endo), 1.26 (m, Hlendo), 1.19 (m, H8a). The position spectra were obtained by using a relaxation time of 4 s between of H4exo was obtained by an NOE experiment using the alcoscans, 64 values of tl, and zero filling to 256 points in fl (lH). hol-OD. I3C NMR: dc (CDCl,) (revised from ref 9) 71.3 (CZ), 41.6 NOE's were obtained by difference spectra, the decoupler offset (Cl), 34.3 (C5), 32.1 (C8), 28.3 (C4), 26.8 (C7), 26.7 (C6), 26.4 (C3). for the reference spectrum being 10 000 Hz. A delay time of 20 The reaction of 3-bromobicyclo[3.2.1]oct-3-en-2-ero-ol(4g) was s was incorporated to ensure complete relaxation, along with a repeated by using deuterium, NaOD, and DzO to give 3,3,4line broadening of 1Hz and an acquisition time of 1.5 s with zero exo-trideuteriobicyclo[3.2.l]octan-2-exo-01 (7) (1.6 g, 63%): 'H filling to 16384 points. Mass spectra were run on an AEI MS902 NMR bD (CHC13) 1.64 (2 D, D3ex0, D4exo), 1.38 (1 D, DSexo), spectrometer. A Hewlett Packard H P 5890A GLC instrument 1.25 (0.34 D, Dlendo); I3C NMR 6c (CDCl,) 71.2 (L.B., C2), 41.5 was used in both analytical and preparative modes. For prepa(Cl), 34.1 (L.B., C5), 32.0 (C8), 27.6 (t, J i v , = ~ 18.7 Hz, C4), 26.7 rative separations 1.5% OV-17 and 1.95% QF-1 on Chromosorb (C7), 26.6 (L.B., C6), 25.8 (m, C3) (L.B. indicates the presence W in a column of 5-mm external diameter and length 3 m was of a small 13C-2H coupling3'). used. Radial chromatography was performed on a Chromatatron Bicyclo[3.2.1]octan-2-exo-ol (1.54 g) was oxidized by chromium (Harrison and Harrison) using Merck type 60 PF254 silica gel. trioxide in acetic acid to give bicyclo[3.2.l]octan-2-one(1.2 g, 79%) endo-Tricycl0[3.2,1.0~~~]0ctane (1). To a suspension of acas a white solid: 'H NMR dH (CDCl3) 2.69 (t, 3J1,7exo = 3J1,8a = tivated palladium on carbon ( 5 % , 80 mg) in pentane was added endo-tricyclo[3.2.1.02~4]oct-6-ene6 (1g) in pentane (20 mL). The mixture was stirred vigorously in a hydrogen atmosphere until (29) (a) Cheng, A. K.; Stothers, J. B. Org. Magn. Reson. 1977,9,355. 1mol of hydrogen had been adsorbed. The mixture was filtered (b) Bicker, R.; Zimmermann, G.; Kessler, H. Chem. Ber. 1978,111,3200. and the solvent removed by distillation through a Vigreux column. (30) (a) Freeman, P. K.; Hutchinson, L. L.; Blazevich, J. N. J. Org. The residual liquid was left in an open flask for a few hours to Chem. 1974, 39, 3606. (b) Freeman, P. K.; Hutchinson, L. L. J. Org. remove the remaining traces of pentane to give endo-tricycloChem. 1980,45, 3191. H3exo and H3endo reassigned.

+

(28) Nyholm, R. S.Proc. Chem. SOC.1961, 273.

(31) A similar line broadening (Le., a small Jisc?~)is observed in the 4-deuteriobicyclo[3.2.l]octan-2-onesystem. Patel, V.; Ragauskas, A. J.; Stothers, J. B. Can. J. Chem. 1986, 64, 1440.

Coxon e t al.

1390 J . Org. Chem., Vol. 54, No. 6, 1989 4.6 Hz, H l ) , 2.44 (m, H5), 2.37 (m, HBexo), 2.19 (4, 2J3endo,3ero

-

16.0 Hz, 3J3en&,4endo = 7.0 Hz, H3endo), ca. 2.0 (HGexo), ca. 1.8 (H4end0, H4ex0, H7ex0, H'lendo), ca. 1.7 (H6endo); 13C NMR, see ref 9. 3,3,4-exo-Trideuteriobicyclo[3.2.l]octan-2-one (9) was prepared as above but by oxidation with chromium trioxide in deuterioacetic acid: 13C NMR 6c (CDC13)210.9 (CZ), 51.1 (Cl), 38.2 (C8), 33.8 (L.B., C5), 31.6 (t, J 1 3 ~ , 2=~ 19.4 Hz, C4), 28.0 (C7), 27.9 (L.B., C6); 2H NMR 6 D (CHC13) 2.33 (0.92 D, DBexo), 2.16 (0.88 D, D3endo), 1.77 (0.28 D, D4endo), 1.69 (0.73 D, D4exo); mass spectrum shows 3% Do, 5% D1, 22% D2,61% D3, 6% D4, 1% D5, 0.6% D,. To 3,3,4-exo-trideuteriobicyclo[3.2.l]octan-2-one (9) (100 mg) in anhydrous ether (10 mL) was added with stirring LiA1H4 (140 mg). After stirring for 3 h, the excess LiAlH4 was destroyed by the careful addition of sodium sulfate decahydrate crystals. Water (10 mL) was added and the aqueous layer extracted with ether. The combined ether extracts were washed with a brine solution and dried over Na2S04. Removal of the solvent at reduced pressure gave a mixture (4:l; 82 mg, 81%) of 3,3,4-exo-trideuteriobicyclo[3.2.l]octan-2-endo-ol (loa) [13C NMR 6c (CDC13) 72.4 (C2), 42.5 (Cl), 37.1 (CS), 33.4 (C5), 30.0 (t, 5 1 = ~ 20.0 ~ Hz, C4), 28.4 (C6), 23.2 (C7, C3, not obsd); 2H NMR 6~ (CHC13) 1.74 (0.85 D), 1.67 (0.3 D, Dlendo), 1.35 (0.90 D), 1.24 (0.75 D)] and 3,3,4-exo-trideuteriobicyclo[3.2.l]octan-2-exo-ol. 3,3,4-exo-Trideuteriobicyclo[3.2.1]octan-2-endo-ol (loa) (94 mg) in dry benzene (5 mL) was added dropwise with vigorous stirring to sodium amide (200 mg) in dry benzene (5 mL). The mixture was heated under reflux for 16 h and the benzene removed by draining the water out of the condenser. Methyl iodide (3.4 g, 1.5 mL) in anhydrous ether (5 mL) was added and the mixture heated under reflux for 8 h. Water (10 mL) was carefully added and the ether layer separated. The aqueous layer was extracted with ether, and the combined ether extracts were washed with a brine solution and dried over MgS04. Removal of the solvent at reduced pressure gave 2-endo-methoxy-3,3,4-exo-trideuteriobicyclo[3.2.l]octane (lob) (48 mg, 47%) as a pale yellow oil: 'H NMR 8~ (CDCl3) 3.30 (OMe), 3.16 ( w h 2 = 6 Hz, H2), 2.35 (4J1,5 = 2.1 Hz, ,J,,& = 6.1 Hz, 3J1,7exo = 6.1 kz, Hl); 2.10 (s), W,,, = 15.8 Hz, H5), 1.85-1.20 (m, 7.2 H); 13CNMR 6c (CDCl,) 81.8 (C2), H 55.6 (OMe), 38.5 (Cl), 37.4 (C8), 34.3 (C5), 30.4 (t, J I ~ C=Z19.1 Hz, C4), 28.8 (C6), 23.9 (C7, C3, not observed); 'H N M R 6~ (CHC1,) 1.78 (0.90 D, D3exo), 1.42 (0.25 D, D4endo), 1.33 (0.71 D, D4exo), 1.23 (0.90 D, D3endo). Reaction of 1 with Mercuric Acetate. To a solution of endo-tricyclo[3.2.1,02~4]octane (100 mg) in anhydrous methanol (4 mL) was added, with stirring, mercuric acetate (300 mg), and the mixture was stirred for 3 h. The mixture was filtered to remove unreacted mercuric acetate, the solvent removed under reduced pressure, and the organomercurial placed under high vacuum for 12 h to remove acetic acid. The residue, a pale green viscous oil, was identified as 4-endo-(acetoxymercurio)-2-endomethoxybicyclo[3.2.1]octane(3c) (275 mg, 95%): 'H NMR 6~ (CDCl,) 3.40 (s, OMe), 3.18 (m, 3J2exo,l = 2.6 Hz, 3J2e,0,3ero = 5.4 Hz, 3J2exo,3endo = 9.8 Hz, HZexo), 2.75 (qt, 3J4exo,3endo = 13.6 Hz, 3J4exa,3exo = 4.7 Hz, ,J4exo,5 = 1.7 Hz, 4J4exo,6exo = 1.6 HZ, J1%g,H4 = 192 Hz, H4), 2.50 (t,3J5,&xo = 3J5,s, = 6.0 Hz, H5), 2.33 (s, Wh,2 = 11.9 Hz, Hl), 2.25 (m, HGexo), 2.00 (s, OAc), 1.9-1.75 (m, H7ex0, H3ex0, HGendo), 1.68 (m, 2J8,,8a = 12.0 Hz, = 3J8a,5 = 5.0 Hz, H8a), 1.55 (m, H3end0, H7endo), 1.38 (d, 2J8a,8s = 12.0 Hz, H8s); 13CNMR 6c (CD30D) 178.4 (OAc), 82.5 ( J I w ,=~310 Hz, C2), 55.7 (OMe), 50.3 ( J l 9 9 ~ ~ , 1 3=c 1634 HZ, C4), 40.8 ( J l s H g , 1 8 C = 66 Hz, C5), 39.3 ( J 1 5 3 ~ ~ 1=3 25 ~ Hz, Cl), 38.8 ( J l s S ~ ~ , l 3=c 310 Hz, C8), 32.7 (Jiwg,nC= 65 Hz, C6), 31.2 ( J I ~ +=~97, I Hz, ~ C3), C 23.8 (J~y+~ not , qobs, C7), 22.9 (OAc). Note: The '%Hg,13C coupling constants are the averages obtained from three separate spectra. Reaction of 1 with Methanol-d/p-ToluenesulfonicAcid. The reaction of endo-tricyclo[3,2.1.02~4]octane (120 mg) with methanol-d was carried out as previously described for the reaction with methanol. The product was isolated after 7 days to give 4-endo- and 6-endo-deuterio-2-endo-methoxybicyclo[3.2.l]octane (3b and 4b) (120 mg, 79%) in the ratio of 62:38 (f4%), respectively: mass spectrum 15% Do, 84% D1, 1% D2;C9H150Drequires M" 141.1264, found M" 141.1269; 2H NMR 6 D (CHC13) 1.41 (D4endo), 1.34 (D6endo). 4-endo-Deuterio-2-endo-methoxybicyclo[3.2.l]octane (3b): 13C NMR 6c (CDC13)81.6 ('221, 55.4

(OMe), 38.4 (Cl), 37.2 (C8), 34.1 (C5), 30.3 (t, J 1 3 C , 1 ~= 20.4 Hz, C4), 28.5 (C6), 25.3 (C3), 23.6 (C7). 6-endo-Deuterio-2-endomethoxybicyclo[3.2.1]octane(4b): 13C NMR 8c (CDCl,) 81.6 (C2); 55.4 (OMe), 38.4 (Cl), 37.2 (C8), 34.1 (C5), 30.7 (C4), 28.2 (t,JiQ,q = 19.9 Hz, C6), 25.4 (C3), 23.5 (C7). Reduction of the Organomercurial 3c with (a) Sodium Borohydride. To the crude 4-endo-(acetoxymercurio)-2-endomethoxybicyclo[3.2.1]octane (275 mg) in methanol (4 mL) was added with stirring aqueous NaOH (5 mL, 1 M), followed by a solution of NaOH (5 mL, 1 M) and sodium borohydride (80 mg). After 30 min the mixture was extracted with pentane, washed with water, and dried over MgS04. The solvent was removed under reduced pressure to give 2-endo-methoxybicyclo[3.2.l]octane (3a) (68 mg, 66%), identical with a sample obtained by the with toluenesulfonic acid reaction of endo-tricyclo[3.2.1.02~4]~ctane in methanol. (b) With Sodium Borodeuteride. T o the crude 4-endo(acetoxymercurio)-2-endo-methoxybicyclo[3.2.l]octane (3c) (100 mg) dissolved in methanol (1mL) was added with stirring aqueous NaOH (1 mL, 1 M) and sodium borodeuteride (20 mg). After 30 min of stirring, the liquid was extracted with pentane, washed with water, and dried over MgS04. The solvent was removed under reduced pressure to give a mixture (4555; 25 mg, 65%) shown to be 4-exo- and 4-endo-deuterio-2-endo-methoxybicyclo[3.2.l]octane, respectively. 'H NMR 6~ (CHC13) 1.40 (D4endo), 1.33 (D4exo). Mass spectrum shows 6% Do, 94% D1. (c) With Sodium Amalgam. Mercury (50 g) was cautiously added dropwise to molten sodium (0.75 g) under Shell Ondina 17 (20 mL). The resulting amalgam was left to cool before being transferred to a mortar where it was broken into small pieces under pentane. To the sodium amalgam (5 g, 1.5%), previously washed with pentane and dried under vacuum for 1h, in NaOD/D20 (2 mL, 2 M) was added the organomercurial 3c (100 mg). The mixture was stirred for 3 h and water (4 mL) was added. The mixture was extracted with pentane, the combined extracts were dried over MgS04, and the solvent was removed under reduced pressure to give 4-endo-deuterio-2-endo-methoxybicyclo[3.2.1]octane (3b) (23 mg, 64%) as an oil: 2H NMR 6D (CHC13) 1.39 (D4endo); 13CNMR 8c (CDCl,) 81.6 (C2), 55.4 (OMe), 38.4 (Cl), 37.2 (C8), 34.1 (C5), 30.3 (t,C4, J 1 3 ~ , 2=~ 20.4 Hz), 28.5 (C6), 25.3 (C3), 23.6 (C7); mass spectrum 6% Do, 94% D1. Reaction of 2 with p-Toluenesulfonic Acid/Methanol at 80 O C . Reaction of exo-tricyclo[3.2.1.02~4]octane(150 mg) with methanol and p-toluenesulfonic acid was carried out as previously The described for the reaction with endo-tricycl0[3.2.1.O~,~]octane. product was isolated after 7 days to give an oil (174 mg, 89%), shown by GLC to contain three compounds (15%, 9%, 76%). Separation of the two major products was achieved by preparative GLC. The minor product was not obtained pure and was not identified. The major product was 2-exo-methoxybicyclo[3.2.l]octane (12a) (76%): 'H NMR 8~ (CDC1,) 3.30 (s, OMe), 3.18 (Wh 2 = 8 Hz, HZendo), 2.38 (S, w h / 2 = 9.5 HZ, HI), 2.14 (Whp = 10.4 H5), 1.84 (d, 2Jh,8.= 10.9 Hz, H8s), 1.7-1.5 (m, Hlexo, HGexo, H7ex0, H'iendo), 1.55-1.42 (m, H3ex0, HSendo, HGendo), 1.38 (m, H4endo), 1.16 (m, H8a); 13C NMR 6c (CDCl,) 80.2 (C2), 55.8 (OMe), 37.9 (Cl), 34.2 (C5), 32.2 (C8), 28.7 (C4), 27.3 (C7), 26.6 (C6), 23.1 (C3). 7-syn-Methyl-2-exo-methoxybicyclo[2.2.l]heptane ( l l a ) (15%): 'H NMR 6~ (CDC13) 3.28 (m, 3J2endo,3endo = 7.5 Hz, 3J2end0,3ex0 = 3.2 Hz, H2endo), 3.26 (s, OMe), 2.10 (br d, 3J1,6ero = 3.1 Hz, H l ) , 1.95 (t, 3J4,5ex0 = 3J4,3exo = 3.0 Hz, H4), 1.8-1.65 (Hlexo, HBendo), 1.65 (m, H7anti), 1.55-1.45 (H5ex0, HGexo), 1.09-1.05 (H5end0, HGendo), 1.05 (d, 3JCH3,7ati = 7.2 Hz, CH3); 13C NMR 6H (CDC13) 85.6 (CZ), 56.0 (OMe), 43.6 (Cl, C7), 40.3 (C4), 36.9 (C3), 28.9 (C5), 26.4 (C6), 12.8 (CH3). An authentic sample of 2-exo-methoxybicyclo[3.2.l]octanewas prepared from bicyclo[3.2.l]octan-2-exo-ol(172 mg) by methylation with sodium amide/methyl iodide to give a yellow oil (124 mg, 65%) consisting of 2-exo-methoxybicyclo[3.2.l]octane (12a) (ca. 56%), 2-endo-methoxybicyclo[3.2.l]octane(ca. 26%), and unreacted starting material. The 'H and 13C NMR data are identical with those obtained above. Reaction of 2 with Mercuric Acetate in Methanol and Subsequent Reduction. T o a solution of exo-tricyclo[3.2.1.02.4]octane(2) (150 mg) in anhydrous methanol (4.5 mL) was added mercuric acetate (580 mg), and the mixture was stirred for 48 h. The mixture was filtered and the solvent removed under

dz,

J. Org. Chem. 1989,54, 1391-1398 reduced pressure. T o the crude mixture was added sodium amalgam (8 g, 1.5% Na/Hg) with NaOD/DzO (5 mL, 1 M), and the reaction was stirred for 3 h. The mixture was diluted with water, the mercury decanted off, and the product extracted into pentane. The combined pentane extracts were dried over MgS04 and filtered and the solvent removed under reduced pressure to yield an oil (142 mg, 73%) shown to be a mixture of 7-syn(deuteriomethyl)-2-exo-methoxybicyclo[2.2.l]heptane (1lb) (43%) (12b) and 4-exo-deuterio-2-exo-methoxybicyclo[3.2.l]octane (48%). Separation was effected by preparative GLC. 4-exoDeuterio-2-exo-methoxybicyclo[3.2.l]octane (12b): 'H NMR 6~ (CDC13) 3.30 (s, OMe), 3.18 (wh/,= 6 Hz, Haendo), 2.38 (s, W,,, = 9.6 Hz, H l ) , 2.14 (9, Wh,Z = 16.2 Hz, H5), 1.84 (d, 'J&,k= 10.9 Hz, H8s), 1.65 (m, HGexo, H7ex0, H'lendo), 1.55-1.42 (HBexo, HSendo, HGendo), 1.38 (d, 3Jqendo,3endo = 9.1 Hz, H4endo), 1.16 (m, 3Jw = 3J%l = 5.7 Hz, H8a); 2H NMR b (CHClJ 1.62 (Dbxo); 13C NMR aC (CDC13) 80.4 (C2), 55.8 (OMe), 38.0 (Cl), 34.2 (C5), = 18.9 Hz, C4), 27.3 (C7), 26.8 (C6); MS, 32.3 (CB), 28.5 (t, J i q , ~ C9H160Drequires M" 141.1264, found M" 141.1264. 7-syn-(Deuteriomethyl)-2-exo-methoxybicyclo[ 2.2.11 heptane (llb): 'H NMR (CDCld 3.28 (m, 3Jamd0= 7.5 Hz, 3 J ~ 0 = 3.2 Hz, Heendo), 3.26 (s, OMe), 2.10 (br d, 3J1,6ex0 = 3.1 Hz, H l ) , 1.95 (t, 3J4,5exo = 3J4,3exo = 3.0 Hz, H4), 1.8-1.65 (H3ex0, = 6.6 Hz, H7anti); 1.55-1.45 (H5ex0, H3endo), 1.65 (t, 3J78,ti,CHzD HGexo), 1.09-1.05 (HBendo, HGendo, CH,D); 2H NMR b (CHCl,) 1.07 (CHzD); 13C NMR 6c (CDCl,) 85.7 (C2), 56.2 (OMe), 43.8 (L.B., C l ) , 43.6 (L.B., C7), 40.5 (L.B., C4), 37.1 (C3), 29.0 (C5),

1391

26.6 (C6), 12.7 (t, J i 3 ~ , 2=~19.1 Hz, CHzD) (L.B. indicates the presence of a small 13C-2H coupling). Reaction of 2 with Methanol-d/p-Toluenesulfonic Acid. Reaction of exo-tricyclo[3.2.1.0z~4]octane(2) (150 mg) with methanol-d was carried out as previously described to give a pale yellow oil (174 mg, 91%), shown to consist of 7-syn-(deuteriomethyl)-2-exo-methoxybicyclo[2.2.1]heptane (11b) (15%) and 4-exo-deuterio-2-exo-methoxybicyclo[3.2.l]octane (12b) (76%). Separation was effected by preparative GLC. The spectral data for these compounds are identical with those obtained from the sodium amalgam reduction of the products from reaction of exo-tricyclo[3.2.1.02~4]~ctane (2) with mercuric acetate/methanol (12b): above. 4-exo-Deuteric-2-exo-methoxybicyclo[3.2.l]octane MS 3% Do, 90% D1, 7% Dz.

Acknowledgment. We acknowledge grants from the Research Committee of the New Zealand Universities Grants Committee. Registry No. 1, 22389-16-8; 2, 13377-46-3; 3a, 85698-98-2;

m 4-endo-3b, 113925-23-8;4-exo-3b, 118760-52-4;3c, 118716-26-0; 4b, 113947-64-1; 7,118716-27-1; 8,2565-97-1; 9, 118722-48-8; loa, 118760-53-5;lob, 118716-28-2; lla, 118716-29-3; l l b , 118716-30-6; 12a, 85698-97-1; 12b, 118722-49-9;endo-tricyclo[3.2.1.02"]oct-6-ene, 3635-94-7; bicyclo[3.2.l]octan-2-exo-ol, 1965-38-4; bicyclo[3.2.l]octan-2-one, 5019-82-9; mercuric acetate, 1600-27-7; methanol-d, 1455-13-6.

Selective Carriers of Ammonium Ions. 1. Synthesis and Template Effect of CsCl and X-ray Structure and Ionophoric Properties of Polyether Crowns Containing l-Methy1-3,5-bis(methylene)-lH-pyrazole Units P. Navarro,*it M. I. Rodriguez-Franco,' C. Foces-Foces,t F. Cano,* and A. Samat* Znstituto de Quimica M6dica (CSZC), Juan de la Cierua, 3, 28006 Madrid, Spain, Znstituto de Quimica F k c a "Rocasolano", Serrano, 119, 28006 Madrid, Spain, and Uniuersit6 d'Aix-Marseille ZZI, LA 126 du CNRS, Rue Henri Poincar6 13397, Marseille Cedex 13, France Received July 15, 1988

A new series of asymmetric crown ethers of cyclic (7, 8) and acyclic structure (9) containing l-methyl-3,5bis(methy1ene)-1H-pyazoleunits have been synthesized. Crowns 7 and 8 having 18 and 36 ring atoms have been obtained in 24% and 25% yields respectively by 1:l and 2 2 cyclization of l-methyl-3,5bis(chloromethyl)-1H-pyazole and disodium tetraethylene glycolate under the template effect of CsC1. When the Cs+ cation was not present in the above reaction, the crowns 7 and 8 were formed in lower yields. The X-ray structure of smaller crown 7 showed a flexible irregular cavity which presents an internal overall twist that changes the pseudo-2-fold axis of symmetry in the neighborhood of the pyrazole ring into a pseudo mirror plane for the rest of it. A 'H and 13C NMR study has shown that in the E ~ ( f o d ) ~ - c r o w7ncomplex, the pyrazolic sp2 nitrogen and the oxygens belonging to the polyether cavity are cooperatively acting as donor sites, the Eu3+being near the pyrazole side least hindered by the methyl group. The transport rates of alkali and ammonium ions are much higher for all polyether ligands 7-9 than for ester crowns 3 and 4 evaluated before. The larger crown, 8, is a selective carrier of NH4+ in relation to K+ and Na+ ions. However, the smaller crown, 7, and its acyclic analogue 9 (which shows a "plateau" selectivity toward NH4+ and alkali ions) are better carriers of K+ ions.

Introduction In view of the important role played by substituted ammonium ions in chemistry and biology, the development of receptor molecules capable of recognizing and transport of such substrates is of special interest.'p2 Macrocyclic polyethers bind primary ammonium ions by anchoring the NH4+group into their circular cavity via three +N-H-O hydrogen bonds3+ and are able to carry organic ammonium cations, in particular, physiologically +

Instituto de Quimica MBdica (CSIC).

* Instituto de Quimica Fisica "Rocasolano". 8 UniversitB

d'Aix-Marseille 111, LA 126 du CNRS. 0022-3263/89/ 1954-1391$01.50/0

active ones.' In general, however, these polyethers bind the alkali cations K+ and Rb+ appreciably stronger than R-NH3+ groups8 whereas the opposite selectivity would (1)Lehn, J. M.Angew. Chem., Znt. Ed. Engl. 1988,27,89. (2) Izatt, R. M.; Bradshaw, J. S.; Nielsen, S. A,; Lamb, J. D.; Christensen, J. J.; Sen, D. Chem. Rev. 1985,85, 271. (3)Stoddart, J. F. Chem. SOC.Reu. 1979, 7, 85; Annu. Rep. Prog. Chem., Sect. B 1983, 353. (4)Cram, D. J.; Cram, J. M. Acc. Chem. Res. 1978, 11, 8. (5)Hayward, R. C. Chem. SOC. Rev. 1983,12, 285. (6)Sutherland, I. 0.Chem. SOC. Rev. 1986,15, 63. (7) Bacon, E.;Jung, L.; Lehn, J. M. J. Chem. Res., Synop. 1980, 136; J. Chem. Res., Miniprint 1980, 1964. (8) Izatt, R.M.; Terry, R. E.; Haymore, B. L.; Hansen, L. D.; Dalley, 1976,98,7620. N. K.; Avondet, A. G.; Christensen,J. J. J.Am. Chem. SOC.

0 1989 American Chemical Society